In-depth reporting and analytical commentary on artificial intelligence regulation. No legal advice.

New York Times would like to pick up where Elon Musk left off, seeks discovery into OpenAI’s departure from non-profit

Context: The AI copyright case brought by the New York Times against OpenAI and Microsoft is still at an early procedural stage, with a focus on pretrial discovery (May 30, 2024 ai fray article). A case with a different focus, which Elon Musk brought in California state court over OpenAI’s departure from a pure non-profit, has meanwhile been voluntarily dismissed (June 11, 2024 CNN.com article).

What’s new: In a letter (PDF) that announces a motion to compel OpenAI to produce certain categories of documents, the New York Times raises, among other things, the issue of OpenAI allegedly refusing to answer its request for information “concerning OpenAI’s transition to a for-profit company.” According to the letter, “OpenAI admitted that while no individual entity has ‘transitioned’ from non-profit to for-profit status, OpenAI did create at least one for-profit entity.” The Times argues that this is relevant in connection with the fair use defense (which ai fray considers the single most critical part of the case).

Direct impact: The letter merely requests a status conference at which OpenAI would like to raise those discovery questions. It remains to be seen what will happen next. The parties have been able to narrow their discovery disputes before.

Wider ramifications: Other parties suing OpenAI in United States courts will probably also try to turn intellectual property or other disputes into a more general debate over the impact of AI and the role OpenAI plays. Judges act as gatekeepers and can limit the extent to which political or philosophical questions are raised in a given context.

In their letter to Judge Sidney H. Stein of the United States District Court for the Southern District of New York, the NYT’s lawyers from Susman Godfrey (a high-power firm that is also suing Microsoft’s AI infrastructure over patents (June 10, 2024 ip fray article)) write the following about OpenAI’s departure from a pure non-profit (references and citations omitted):

OpenAI Refuses to Produce Documents About Its Formation of a For-Profit Entity.

RFP 11 seeks documents “concerning OpenAI’s transition to a for-profit company.” Initially, OpenAI sought to evade this RFP through word games, arguing that “no OpenAI entity has undergone” any “transition.” But when pressed during the meet and confer, OpenAI admitted that while no individual entity has “transitioned” from non-profit to for-profit status, OpenAI did create at least one for-profit entity. OpenAI nonetheless continues to refuse to produce documents responsive to this request.

This Court should order OpenAI to produce documents concerning its decision to form one or more for-profit entities. OpenAI’s commercial designs bear on Defendants’ anticipated fair use arguments, which include consideration of the “purpose and character of the use” of copyrighted works. OpenAI has suggested that discovery into this topic is unnecessary because it will not dispute that it charges fees to certain users of ChatGPT, but that does not come close to providing The Times the discovery it needs on this aspect of fair use. OpenAI touts that its work is intended to benefit all of humanity and it will undoubtedly press that narrative. That OpenAI chose to create a for-profit entity through which it conducts most of its operations and solicited massive investments from Microsoft and others directly undermines that story. The Times is entitled to discovery on that choice.

ai fray believes that the NYT may very well be entitled to discovery into OpenAI’s departure from a non-profit entity. Fair use is a question to be decided by the judge, so there is no risk of jury confusion at that procedural stage (summary judgment).

If, however, the NYT surmounts all the hurdles between its complaint and a jury trial (of which fair use is the most difficult one), then there should be a limit as to the extent to which the NYT can talk about OpenAI’s history. There comes a point where the probative value of any of those stories would be outweighed by jury confusion.

The NYT’s lawyers presumably hope to find documents from OpenAI’s history that will make a psychological impact on jurors. A purely rational analysis of the fair use defense is all about whether and how OpenAI has been using any of the copyrighted material as opposed to what could have been or might have been if OpenAI had remained a pure non-profit.